CS355 Sylabus

Final Notes

We took a "nested-if" approach to decode an assembler instruction, in other words, we wrote a nested-if program to find the sequence of mirco-instructions that perform the instruction/command encoded in the assembler code
The ultimate goal of the decoding process is to determine the segment of micro-porgram that contains micro-instruction codes that instruct the CPU to "behave" according to the assembler instruction.
In the "nested-if" approach (the one used in the lecture notes), the datapath execute many (4 to be precise) branching micro-instructions to reach the segment of micro-porgram that contains micro-instruction codes that instruct the CPU to "behave" according to the assembler instruction.
There is a more efficient way to find this segment of micro-porgram...
A more efficient deciding approach is the "table-driven" approach.
You must construct a table with addresses in the micro-program that are the starting addresses of program segments that instruct the CPU to "behave" according to certain assembler instructions.
In our example, the micro-program segment to execute a load assembler instruction starts at address 6, while the micro-program segment to execute a store assembler instruction starts at address 9, etc.
We would have the following "branch address" table:
When a load instruction is fetched, we can use the table to quickly branch to the segment of micro-program that executes load instructions using the branch table to obtain the MPC address:
Similarly, when a store instruction is fetched, we can use obtain the address of the segment of micro-program that executes store instructions: